Our attorney is on a lawsuit with a pro se litigant who is using chat gpt for his motions. In his depo of the defendant he asked him what AI he uses and if he can give him his list of prompts he used. His hope is the guy admits to his plaintiffs claims in the prompts.

Think the pro se guy tries to claim privilege?

Is it privileged?

Attorneys using chatgpt for their case would be privileged. An individual using chatgpt themselves is not. It's no different than searching the web.

There is a recent case where AI data is not privileged if it is a public platform where your inputs/documents are used for machine learning. My understanding is that if your data is not use to teach the program, then it is privileged.

This is the TEXAS law on Attorney Work Product privilege form the TSC from a 2017 case that was about attorney/client billing records:

Quote:

---

"The work product privilege is broader than the attorney-client privilege." The discovery rules define "work product" as: (1) material prepared or mental impressions developed in anticipation of litigation or for trial by or for a party or a party's representatives . . . ; or (2) a communication made in anticipation of litigation or for trial between a party and the party's representatives or among a party's representatives, including the party's attorneys [and] consultants.

---

I'm not going to do a full research project and try to answer the specific question, but, absent intervening case law regarding AI, I would have (based on this) a very difficult time arguing that such work the OP is discussing is NOT covered under the Work Product privilege. I don't think the court would have to use mental gymnastics to find, IF NEEDED (and I'm not sure it is), that "consultants" includes AI just like an email would be like a written correspondence. I personally think it is easily covered in (1), but I guess some might disagree.

But again, whether this has been litigated or not in Texas courts and there is a different ruling, I don't know.

Haven't read it yet, but assume this article has some good info.

Attorney Intelligence?

itsyourboypookie said:

---

Our attorney is on a lawsuit with a pro se litigant who is using chat gpt for his motions. In his depo of the defendant he asked him what AI he uses and if he can give him his list of prompts he used. His hope is the guy admits to his plaintiffs claims in the prompts.

Think the pro se guy tries to claim privilege?

Is it privileged?

---

I don't know, but he'll definitely try to claim it's privileged. It would at little dumb not to and just give it up without a fight.

IIIHorn said:

---

Attorney Intelligence?

---

Contradiction in terms, like jumbo shrimp.

I.e. if you are working with an attorney on a legal matter, you tell the attorney what to run through AI? I suppose the grey area is what happens if you are self representing?

Claude! said:

---

IIIHorn said:

---

Attorney Intelligence?

---

Contradiction in terms, like jumbo shrimp.

---

Doh!

Never ask to see their briefs.

This is correct.

What's the test on that?

If I Google something like "legal strategies to get out of paying money if I hit someone while running a red light" is that considered work product in preparation for litigation? Or better, "how to win a criminal trial if I'm guilty of murder but they'll never find where I hid the body?"

Claude! said:

---

IIIHorn said:

---

Attorney Intelligence?

---

Contradiction in terms, like jumbo shrimp.

---

Indicative of shellfishness?

Maybe I'm practicing a different type of law.

Queso1 said:

---

Maybe I'm practicing a different type of law.

---

"Practicing"?

I like lawyers who actually know how to do things, not ones who practice doing things.

Interesting development on a related note.

Quote:

---

Quote:

---

Thomson Reuters and Anthropic Expand Partnership to Connect Claude with CoCounsel Legal
New MCP integration brings CoCounsel Legal into Claude workflows to help legal work meet fiduciary-grade standards

---

---

https://www.thomsonreuters.com/en/press-releases/2026/may/thomson-reuters-and-anthropic-expand-partnership-to-connect-claude-with-cocounsel-legal

Ok. That's pretty funny.

AustinAg2K said:

---

Attorneys using chatgpt for their case would be privileged. An individual using chatgpt themselves is not. It's no different than searching the web.

---

Absolutely wrong. The web doesn't give you customized advice for your case at anywhere near the speed ChatGPT does.

TexAgs91 said:

---

AustinAg2K said:

---

Attorneys using chatgpt for their case would be privileged. An individual using chatgpt themselves is not. It's no different than searching the web.

---

Absolutely wrong. The web doesn't give you customized advice for your case at anywhere near the speed ChatGPT does.

---

How does it being faster change the analysis of whether its privileged?

It is a Dallas guy that kicked off most of this discussion. He was just found guilty of lots of federal crimes including:

" securities fraud, wire fraud, conspiracy to commit securities fraud and wire fraud, and false statements to auditors in connection with a scheme to fraudulently extract more than $150 million from GWG."

His home computers were seized during a raid and analysis showed him using Claude to test out defense strategies. He then took some of his thoughts into the his attorney. Counsel tried to argue the evidence should be privileged. The courts said no sir.


https://www.chapman.com/publication-federal-court-rules-that-ai-generated-documents-are-not-protected-by-privilege

Quote:

---

Heppner's counsel informed the government that Heppner created the AI Documents before his arrest and later shared them with his attorneys, and that some of the information Heppner input into Claude he had learned from his counsel.4 Heppner asserted that the AI Documents were privileged and requested their segregation from the Government attorneys, a request to which the Government agreed pending a privilege determination by the Court.5

The Government made a number of arguments as to why the attorney-client privilege did not apply to the AI Documents:

• The AI Documents were not communications between Heppner and his counsel. Rather, they were made between Heppner and a public AI tool that is not an attorney. The Government analogized Heppner's use of Claude to him asking friends for input on legal matters, an act that does not create privileged communications.8
• The AI Documents were not created for the purpose of obtaining legal advice from counsel. Specifically, Claude's "Constitution," terms of service, and other public materials expressly disclaim Claude's ability to give legal advice and instead suggest that the user consult with a qualified lawyer.9
• The AI Documents were not confidential. Rather, Heppner voluntarily shared his prompts with a third-party commercial AI platform that is publicly accessible. Claude's Privacy Policy explicitly advises users that it collects data on prompts and outputs, uses this data to train its AI tool, and may disclose this data to governmental regulatory authorities and third parties.10

• Heppner's subsequent transmission of the AI Documents to his attorneys did not retroactively cloak them with privilege. Well-settled law holds that sending non-privileged documents to counsel does not make them privileged.11 It is equally well-established that while a communication to one's attorney may be privileged, the underlying information or facts in that communication are not.12

The Court's Ruling
The Court orally granted the Government's motion from the bench on February 10, 2026, holding that Heppner failed to meet his burden of establishing both attorney-client privilege and the work product privilege over the AI Documents.14
Takeaways and Implications

While this ruling is based on fundamental and well-established rules regarding attorney-client privilege and the work product doctrine, and thus is not necessarily surprising, it nevertheless has potentially far-reaching consequences for non-lawyer individuals and organizations who use public AI tools to research legal matters:

• Communications with public AI tools may not be privileged. Communications with public AI tools may not satisfy the requirements of the attorney-client privilege because AI tools are not attorneys, do not provide legal advice, and inputs to them are not confidential. Further, later sending the AI-generated results to a lawyer will not retroactively cloak them with privilege.
• Public AI tool privacy policies are important. Courts may examine whether AI tool privacy policies permit disclosure of user data to third parties and governmental authorities when evaluating confidentiality claims so such policies should be carefully reviewed prior to use of the AI tool.
• Sharing privileged information with a public AI tool might waive privilege. Taking the ruling a step further, it is reasonable to also conclude that sharing confidential attorney-client communications with a public AI tool might waive any privilege that could otherwise attach to those communications.

---

from your cited article:

Quote:

---

The Government also argued that the AI Documents were not shielded by the work-product privilege, which "provides qualified protection for materials prepared by or at the behest of counsel in anticipation of litigation or for trial."13 Heppner admittedly created the AI Documents on his own initiative, not at counsel's direction, and thus should not be able to later claim that he prepared the documents at the behest of counsel in anticipation of trial, as required to assert the work product privilege.

---

in many jurisdictions, including Texas, materials do not have to be "prepared by or at the behest of counsel" to potentially qualify as work product

Quote:

---

in many jurisdictions, including Texas, materials do not have to be "prepared by or at the behest of counsel" to potentially qualify as work product

---

And in an ironic twist, just saying their attorney directed them to do something can then be used as a waiver of all privilege.

I was a party to a lawsuit years and years ago and was being deposed. Questions put to me about why I had done X, Y, and Z. I did those because one of my attorneys had asked me to do that as part of their advice. Each time I had to stop and ask (on the record) counsel sitting next to me if I could even answer that or not. Response was no I could not answer that. I understood his concern was about waiving privilege. And the circumstances were such that the attorney with me at the depo was not the attorney who had given me the advice of what to do. He even said on the record, "I wish you could answer that question but you can't."

So how would that square with waiver issues? I liken it to a client who goes to the Supreme Court law library, does research and takes notes to give back to the attorney of case cites. (Not unusual as clients try to save some money.) Is that work product? I lean towards yes but then one also the waiver issue by even saying that.

Quote:

---

a client who goes to the Supreme Court law library, does research and takes notes to give back to the attorney of case cites. (Not unusual as clients try to save some money.) Is that work product?

---

texas rule:

Quote:

---

Work product defined. Work product comprises:
(1) material prepared or mental impressions developed in anticipation of litigation or for trial by or for a party or a party's representatives, including the party's attorneys, consultants, sureties, indemnitors, insurers, employees, or agents; or

---

mental impressions developed for trial by a party.

clearly work product in texas.

What I thought. Thanks.

So, why does it make a difference if the info from comes from a publicly available source of a book with caselaw versus the publicly available source of AI?

(P.S. Glad I am retired and don't have to deal with this stuff.)

for work product, it may not make a difference. whether it AI or the public library, could be the same result.

the NY case (heppner) did not discuss what bearing the material coming from AI platform had on the determination of if it was work product. the sole basis was that it was not created by or at the behest of an attorney.



there has not been any case (at least that I know of) that does a deep analysis of AI and its application to work product. but there will be soon.

ErnestEndeavor said:

---

What's the test on that?

If I Google something like "legal strategies to get out of paying money if I hit someone while running a red light" is that considered work product in preparation for litigation? Or better, "how to win a criminal trial if I'm guilty of murder but they'll never find where I hid the body?"

---

the Prosecution absolutely used Google searches of

'best prison for women inmates" and "how to delete information from FBI" in a recent murder trial in Utah

P.H. Dexippus said:

---

There is a recent case where AI data is not privileged if it is a public platform where your inputs/documents are used for machine learning. My understanding is that if your data is not use to teach the program, then it is privileged.

---

I was talking about this with my paralegal gf recently. Any chance you have a link to a news article about it?

Less Evil Hank Scorpio said:

---

P.H. Dexippus said:

---

There is a recent case where AI data is not privileged if it is a public platform where your inputs/documents are used for machine learning. My understanding is that if your data is not use to teach the program, then it is privileged.

---

I was talking about this with my paralegal gf recently. Any chance you have a link to a news article about it?

---

(Complete joke here, Just call it a Mom Joke)

Can't she ask Chat for that?

Why would this be any different than the rules governing internet searches? From a legal standpoint it doesn't seem like there would be any meaningful distinction?

Aren't all searches used to train the program? While I wouldn't, communication with a client on SM is probably mined for data for training the algorithm. Google mail mines data…same thing. Google searches do as well.

BMX Bandit said:

---

So the answer is . . . . . . .


It depends

---

This is like being Lawyer Rick Rolled

From Grok: Doesn't Depends

Yes, a prosecutor can likely subpoena prompts, transcripts, or related data from Grok/xAI in a criminal case, and there is no attorney-client privilege or similar protection shielding them.

Why This Is Possible

xAI complies with valid legal process: xAI maintains a dedicated law enforcement portal for subpoenas, warrants, court orders, and preservation requests. Like other U.S. tech companies, xAI discloses user data (including conversation content) when required by law. Their privacy policy explicitly states they retain data for legal obligations and will respond to subpoenas or court orders.

No attorney-client privilege: Conversations with Grok (or similar AIs like ChatGPT/Claude) are not privileged. A recent federal ruling in United States v. Heppner (S.D.N.Y., 2026) held that AI-generated documents and underlying prompts are not protected by attorney-client privilege or work-product doctrine. The defendant used Claude to prepare defense materials; the court ruled they were discoverable because:
The AI is not a lawyer or the lawyer's agent.
There is no reasonable expectation of confidentiality due to the platform's privacy policy and data practices.
The materials were not prepared under the direct supervision of counsel for litigation. This logic applies directly to Grok. Using it for motions or legal strategy creates no special shield.

Data retention and access: xAI retains conversations unless deleted (and even then, often for up to 30 days or longer for legal/safety reasons). Private chats delete faster but can still be held for legal compliance. Prosecutors can:
Subpoena xAI directly for server-side records (prompts, responses, metadata like timestamps/IPs).

Obtain data from your devices/accounts via search warrant or subpoena (e.g., exported chats, screenshots, or browser history).

Practical Risks in Your Scenario

Pro se defendant using Grok for motions: Everything you input (e.g., case facts, strategy, draft language) could become evidence. Prosecutors have successfully used AI chat histories to show intent, planning, or inconsistencies.

Discovery in criminal cases: Prosecutors routinely subpoena tech companies for user data. AI chats are treated like Google searches or emailsdiscoverable with proper process.

No "pro se exception": Forgoing a lawyer doesn't create extra protections; it removes potential work-product arguments that might apply if counsel directed the AI use.

Recommendations
Treat Grok (or any public AI) as non-confidential. Do not input sensitive case details, admissions, or strategies you wouldn't want the prosecutor to read. For real legal help, consult a licensed attorneyAI is a tool, not a substitute, and using it this way can harm your case.

Laws and policies can evolve, but current U.S. precedent and xAI's practices make subpoena access straightforward. If facing charges, speak to a defense lawyer immediately rather than relying solely on AI.

exhibit A in why you should not listen to grok for legal advice